Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com
Message-ID: <4151FEB1.2030004@hones.org.uk>
Date: Wed, 22 Sep 2004 23:37:37 +0100
From: Cliff Hones <cliff AT hones DOT org DOT uk>
User-Agent: Mozilla Thunderbird 0.7.2 (Windows/20040707)
MIME-Version: 1.0
To: Cygwin List <cygwin AT cygwin DOT com>
Subject: Re: Lot of undefined symbols at link time, even with -l option on good libraries
References: <4151B11A DOT 2020209 AT atosorigin DOT com> <NUTMEGFCHOzyfyz5b6K00000028 AT NUTMEG DOT CAM DOT ARTIMI DOT COM> <6 DOT 1 DOT 0 DOT 6 DOT 0 DOT 20040922133349 DOT 03eb1438 AT pop DOT prospeed DOT net>
In-Reply-To: <6.1.0.6.0.20040922133349.03eb1438@pop.prospeed.net>
X-Enigmail-Version: 0.84.2.0
X-Enigmail-Supports: pgp-inline, pgp-mime
Content-Type: text/plain; charset=ISO-8859-1
X-IsSubscribed: yes
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by delorie.com id i8MMbnJg027592

Larry Hall wrote:
> At 01:14 PM 9/22/2004, you wrote:
> 
>>>-----Original Message-----
>>>From: Frédéric ORMANCEY
>>>Sent: 22 September 2004 18:07
>>
>>>I did it, but it don't work !
>>
>> Ah, ok, that was not clear from your last message!
>>
>>
>>>The --enable-stdcall-fixup option suppress one warning on top 
>>>of linker output trace :
>>>    AVERTISSEMENT: résolution de _GetModuleHandleA par un 
>>>lien vers _GetModuleHandleA AT 4
>>>and unfortunatly leave ALL other errors with no effect on it.
>>
>> I see.  Hmm.  That's very strange; it should affect all the errors - that
>>is, if they're all the same kind of error.  OTOH it could be that you've
>>gotten the calling conventions wrong with the other pragmas, so that's why
>>--enable-stdcall-fixup doesn't work for them; there must be some different
>>reason why the errors are caused, since they behave differently.
>>
>> You might care to experiment with the "--enable-auto-import" and, failing
>>that, "--enable-extra-pe-debug" options as well.
> 
> Or Frédéric could try specifying the link symbol he wants in the pragma
> as suggested by an earlier poster (sorry, the thread is broken in the 
> archives and I didn't put any effort into wading through the messages 
> to trace it all back to that message).

That was me - but it seems Dave Korn's post broke the thread.  But isn't
it easy enough to sort messages on subject?

Anyway, I tried a noddy example, as follows...

Cliff AT enigma ~
$ cat imp.adb
procedure imp is
  procedure Doit(A : integer; B : boolean; C : character);
  pragma Import (Stdcall, Doit, "FormatMessageA");
begin
  Doit(1, true, 'x');
end;

Cliff AT enigma ~
$ gnat compile imp.adb
gcc -c imp.adb

Cliff AT enigma ~
$ nm imp.o
00000000 b .bss
00000000 d .data
00000000 t .text
         U _FormatMessageA AT 12
00000000 T __ada_imp

So Gnat is generating a stdcall-style linker symbol with @12 (the
parameter block size) correctly appended.  If Frédéric is not seeing
this with stdcall, it is probably an indication that the specification
used for Doit is wrong - ie it hasn't been defined to take any parameters.

If you change the pragma line to
    pragma Import (Stdcall, Doit, Link_Name => "FormatMessageA");
the symbol is generated as
         U FormatMessageA AT 12

so you can generate it without the leading underscore, but with the
stdcall postfix, if you wish.

[Note - the above is just an example - the *real* FormatMessageA should
take 7 words (28 bytes) of parameters, and the Ada spec should match the
kernel32 definition.]


-- Cliff


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/